Principal Product Manager, Data Intelligence and AI Governance
Adobe
San Jose, CA, US
Onsite
2026-06-24
Announced salary
$134k–$254k
Low
$141K
Median
$198K
High
$270K
Market in San Jose · BLS OEWS 2025
Job description
The Opportunity
Adobe's data platform is built on a set of deeply connected infrastructure products that together form a vertical data intelligence stack. Raw signals are governed and routed at the point of ingestion, enriched and catalogued into curated enterprise definitions, and ultimately transformed into retrieval\-ready, agent\-optimized knowledge assets. AI agents, self\-service analytics, and enterprise decision systems consume the output of this stack every day.
The Principal PM, Data Intelligence \& AI Governance owns the strategic direction and implementation for the cross\-cutting concerns that make this stack trustworthy and agent\-ready: metadata strategy, governance frameworks, data lineage, and the readiness of enterprise data assets for AI consumption. This is not a feature PM role. It is a platform\-level position responsible for the quality, coherence, and trustworthiness of data as it moves through each layer of the platform — and for the unified operator experience that makes that trust visible!
What you will own
Metadata Strategy
Define and drive Adobe's enterprise metadata model — what gets catalogued, how it is structured, what it means, and how it stays current across systems.
Own the product roadmap for metadata enrichment, normalization, and publication — including benchmark definitions, event schemas, data job lineage, and entity relationships.
Partner with the Metadata System PM to translate the metadata strategy into prioritized product features and a coherent data model.
Establish metadata standards that external teams (product analytics, ML, BI) can build on with confidence.
Governance \& Agent Readiness
Own the product definition of 'agent\-ready data' — the governance, freshness, lineage, and trust properties
**Define the cross\-product impact analysis capability:** surfacing what breaks across the full stack when an event schema changes, a benchmark definition is updated, or a kn